Review Metallocene and related catalysts for oleﬁn, alkyne …...Review Metallocene and related catalysts for oleﬁn, alkyne and silane dimerization and oligomerization Christoph

Coordination Chemistry Reviews 250 (2006) 66–94

Review

Metallocene and related catalysts for olefin, alkyne and silanedimerization and oligomerization

Christoph Janiak∗

Institut für Anorganische und Analytische Chemie, Universität Freiburg, Albertstr. 21, D-79104 Freiburg, Germany

Received 8 December 2004; accepted 17 February 2005Available online 31 May 2005

Dedicated to Prof. Helmut Alt on the occasion of his 60th birthday

Contents

1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 672. Oligomer analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

2.1. Molecular weight determinations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692.1.1. Gel permeation/size exclusion chromatography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

0

0

3

512244

6667

2.1.2. 1H NMR spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702.1.3. Gas chromatography, GC and GC/mass spectrometry, GC/MS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702.1.4. Viscosity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.1.5. Cryoscopy and vapor pressure osmometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702.1.6. Raman spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

2.2. End group analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.2.1. 1H and13C NMR spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702.2.2. MALDI–TOF MS (matrix assisted laser desorption ionization–time of flight mass spectrometry). . . . . . . . . . . . . . . . 70

3. Metallocene complexes for olefin, alkyne and silane oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 733.1. Ethene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.2. Ethene/�-olefin co-oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753.3. Propene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.4. 1-Butene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.5. 1-Pentene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.6. 1-Hexene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.7. 1-Heptene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.8. 1-Octene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.9. Branched�-olefin oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 843.10. Cyclic olefin oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

3.10.1. Cyclopentene. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.10.2. Norbornene. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.11. �,�-Diene oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.12. Alkyne oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.13. Dehydrocoupling/dehydrooligomerization of hydrosilanes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Abbreviations: nBu,n-butyl, C4H9; iBu, iso-butyl, (CH3)2CHCH2–; tBu, tert-butyl, Me3C–; Cp, cyclopentadienyl, general; Et, ethyl, C2H5; Flu′, fluorenylconnected to another ligand or substituent, C13H8; Ind, unsubstituted indenyl, C9H7; Ind′, indenyl connected to one other ligand or substituent, C9H6, 1-indenylunless stated otherwise; Ind′′, indenyl connected to two other ligands or substituents, C9H5; Ind′ ′ ′, indenyl connected to three other ligands or substituents,C9H4; Ind′H4, tetrahydroindenyl connected to another ligand or substituent, C9H10; MAO, methylalumoxane; Me, methyl, CH3–; Ph, phenyl, C6H5–; nPr,n-propyl, C3H7; iPr, iso-propyl, (CH3)2CH–; THF, tetrahydrofuran

∗ Tel.: +49 7612036127; fax: +49 7612036147.E-mail address: [email protected].

0010-8545/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.ccr.2005.02.016

C. Janiak / Coordination Chemistry Reviews 250 (2006) 66–94 67

4. Complexes related to metallocenes for olefin oligomerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 904.1. Sandwich complexes with cyclic�-perimeters other than cyclopentadienyl. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 904.2. Half-sandwich complexes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Abstract

This review summarizes the use of metallocene complexes and related compounds as catalysts in the dimerization or oligomerization ofolefins (alkenes) or terminal acetylenes (alkynes) and in the dehydrocoupling/dehydrooligomerization of silanes. Metallocene complexesof group-III metals (scandocenes, yttrocenes, lanthanocenes), lanthanoids (neodymocenes) and group-IV metals (titanocenes, zirconocenes,hafnocenes) have been utilized in the selective (co-/hydro-)oligomerization of ethene, of the�-olefins propene, 1-butene, 1-pentene, 1-hexene, 1-heptene, and 1-octene, of branched olefins, e.g. methyl-butenes, methyl-pentenes and styrene, of cycloolefins, e.g. cyclopenteneand norbornene and of�, �-dienes, e.g. 1,5-hexadiene and 1,7-octadiene. Group-III metallocenes are often active in the C–C coupling withouta cocatalyst; group-IV metallocenes require the help of a cocatalyst, such as methylalumoxane, MAO, aluminum alkyls, e.g. AliBu3, orperfluorated boranes, e.g. B(C6F5)3. The actinoid metallocenes (C5M5)2AnMe2 with An = thorium, uranium allow for the dimerization andoligomerization of terminal acetylenes. The dehydrooligomerization of (hydro)silanes is typically achieved by group-IV metallocene chloridestogether withn-butyl lithium. Also included in this review are related sandwich and half-sandwich (mono-cyclopentadienyl) complexes usedfor olefin oligomerization. The related sandwich complexes feature phospholyl, boratabenzene or carboranate ligands. Methods of oligo–olefinanalyses by gel permeation chromatography (GPC),1H NMR spectroscopy, gas chromatography (GC) or viscosity measurements for molecularweight determinations and by1H and13C NMR spectroscopy or MALDI–TOF mass spectrometry for end group structure determinationsare summarized. Possible applications of olefin oligomers, in particular oligopropenes are presented. The functionality of a double bond atthe end of each chain (for further modifications) together with the product homogeneity are the advantages of oligomers from metallocenecatalysis. In addition, olefin oligomerization is used to study mechanistic aspects and to obtain a better insight into the reaction mechanism ofmetallocene polymerization catalysis because of the homogeneity of the reaction mixture and because certain mechanistic aspects are easier

catalysts” refer to zirconocene complexes, which are in the

inen-ms

eseitera-f

eb-p-

l-

such

ewne,

ne,and 1-hexene[37]. Hence, scandocene[38], yttrocene,

to investigate in oligomeric products than in high-molar-mass polymers.© 2005 Elsevier B.V. All rights reserved.

Keywords: Metallocene catalysts; Olefins; Alkynes; Silanes; Oligomerization; Hydrooligomerization

1. Introduction

The use of olefin oligomers as intermediates for spe-cialty chemicals, versatile feedstocks and building blocksdrives the interest in the catalytic oligomerization[1,2].Ethene oligomers, in particular higher linear�-olefins ofchain lengths C4–C18 are well established and nickel chelatecomplexes used in the SHELL-higher-olefin-process (SHOP)feature prominently in their manufacturing[1,3,4]. A varietyof metals and methods are available for the oligomerization of�-olefins[4–6], but feasible processes involving well-definedcatalysts are rare[7].

The formation of olefin chains is categorized as follows:dimerization whenn = 2, oligomerization when 2 100[1].

The term “metallocene catalyst” typically refers to thebent metallocenes of the group-IV transition metals tita-nium, zirconium and hafnium. In particular, “metallocene

center of academic and industrial attention; see reviews[8–21]. The titanocene catalysts are unstable at convtional polymerization temperatures and the hafnium systeare too expensive. From the beginning of the 1990’s thbis(cyclopentadienyl)group-IV metal complexes (single-scatalysts) were introduced in industry as a new genetion of Ziegler-Natta catalysts for the polymerization oolefins [22–30]. Neutral group-III (Sc, Y, La) and lan-thanoid metallocene complexes of the general type Cp2MRand [Cp2MH]2 are also capable of polymerizing ethen[31–34]. They are polymerization active even in the asence of a cocatalyst. In their monomeric form the grouIII and lanthanoid metallocenes areiso-d-electronic (d0, notcounting the f-electrons) with the active group-IV metalocene cation [Cp2ZrR]+. Group-III and lanthanoid metal-locene catalysts show the same mechanistic features,as olefin insertion,�-H and �-Me elimination and M Rbond hydrogenolysis as the zirconocene catalysts. Nstudies in this direction also involve divalent samaroceCp2Sm [35] as well asansa-Cp2Sm(THF)2 [36]. A single-component chiral organoyttriumansa-metallocene catalyst[Me2Si(C5H2-2-SiMe3-4-tBu)2YH]2 was found to slowlypolymerize propene (97% isotactic), 1-butene, 1-pente

68 C. Janiak / Coordination Chemistry Reviews 250 (2006) 66–94

lanthanocene and lanthanoid metallocene complexes arealso efficient catalysts for the (hydro-)dimerization and(hydro-)oligomerization of�-olefins, isobutene and alkynes(see below).

In the beginning of metallocene catalysis the observationof oily, oligomeric or low molar-mass products from propenepolymerizations with zirconocene catalysts was generally re-garded as unfortunate[39–42]and to some extent still is. Itis now more and more recognized that metallocene catalystscan be used effectively for the directed oligomerization of�-olefins[7,43–58]and other monomers. It seems desirableto use the fascinating features of metallocene catalysts in or-ganic synthesis. The key to obtain oligomers is the drasticreduction of the rate of propagation versus the rate of ter-mination. This diverts the catalytic carbon–carbon couplingprocess from a rapid polymerization to e.g. the syntheticallyuseful dimerization of olefins.

The correlation between the rate constants and the molec-ular weight of the oligomer can be expressed in math-ematical terms[59]: we assume for simplicity only onechain transfer mechanism whose rate is independent ofany species concentration other than the active complex(cat*), e.g. �-hydrogen transfer to metal with rate(�-Helim.) =k�-Hccat*. Then the following equations can be de-rived, usingcM = concentration of monomer;mM = molecularweight of monomer;c = concentration of active complex;k forse

-

- greea

-

-

y

-

-

eo teo

cang ups,p sac Onew , hy-

drogenolysis of the MR bond as an additional terminationreaction. But this has the disadvantage of yielding unfunc-tionalized, saturated hydro-oligomers.

For oligomers at large, a number of potential applica-tions are possible, such as long chain 1-alkenes for surfac-tants, clean diesel fuel alkanes[60], octane boosters and newmonomers[61–64]. In addition, a variety of functionalizationreactions with double-bond end groups in oligomers lead toorganic specialities with potential applications as adhesives[65], blend compatilizers[47,48,66], fragrances, syntheticlubricants[55,67,68], additives for fuels[68] or in the pa-per and leather industry[7,69]. Also, �-olefin oligomers orderivatives thereof can be used as (macro-)monomeric build-ing blocks for novel graft copolymers containing oligo-olefinside chains. The key is the chain end functionalization ofthe usually vinylidene-terminated oligopropenes via doublebond conversion reactions. Succinic anhydride-terminatedoligopropenes are obtained by ene-type addition of maleica actsw -b-p aree nesa ylatep owno ion,t xy-t tonep lockc go-p danto

haint tageso efino nd too etal-l neityo merp mericp poly-m

ligo-o bya cen-t tedt ationa sseso ed

cat*

(propagation)= rate constant for chain propagation and,implicity here,kt = ktermination= k�-H = rate constant for�-Hlimination.

probability for chain growth:p = kpcMccat*/(kpcMccat* +ktccat*) = kpcM/(kpcM + kt)mole fraction of oligomers which have reached the deof oligomerizationi: xi = pi–1(1− p) thereby assumingconstant growth probabilityp. The term (1− p) is the probability for chain transfer with 1–p = kt/(kpcM + kt)number average degree of polymerizationPn =n∑

i=1ixi = (1 − p)

n∑i=1

ip(i−1) ≈ 1 + p + p2 + · · · + pn−1;Pn = p

n−1p−1 = 1−p

n

1−p ≈ 11−p for n becoming sufficientllarge andp < 1 so thatpn → 0.Pn = 11−p =

kpcM+ktkt

= kpcMkt

+ 1number average molecular weightMn = mM

n∑i=1

ixi =

mMPn = mM(

kpcMkt

+ 1)

.

Thus the molecular weightMn is proportional to the ratf propagation,kpcM and inversely proportional to the raf chain transfer,kt.

With metallocenes the chain-termination reactionsive oligomers with unsaturated, double-bond end groredominantly of the vinylidene type (1). In special caselso a vinyl/allyl double-bond (2) or a 2-butenyl group (3)an be formed (for their originating reaction see below).ay to enforce oligomers is to use hydrogenation, that is

nhydride. The anhydride-terminated oligopropene reith amin-terminated polyamide-6,6 to yield polypropeneolyamide-6,6-b-polypropene triblock copolymers whichfficient dispersing agents. Thiol-terminated oligopropere chain transfer reagents in radical methylmethacrolymerization. Acrylic monomers and styrene are grnto the thiol end group via a chain transfer react

hereby producing a family of block copolymers. Hydroerminated oligopropenes are useful initiators in caprolacolymerization to form poly(propene-b-caprolactone) bopolymers. Homo- and copolymerization of acrylic oliropene yield novel classes of graft copolymers with penligopropene chains[46–49](Scheme 1).

The functionality of a double bond at the end of each cogether with the product homogeneity are the advanf oligomers from metallocene catalysis. In addition, olligomerization is used to study mechanistic aspects abtain a better insight into the reaction mechanism of m

ocene polymerization catalysis because of the homogef the reaction mixture (no heterogenation through polyrecipitation) and because in some aspects the oligoroducts are easier to investigate than high-molar-massers[70–84].sThe average molecular weight of the produced o

lefins for a given catalyst system can be controlleddjustment of reaction temperature and monomer con

ration, with the molecular weight being inversely correlao temperature, while an increase in monomer concentrlso increases the molecular weight. The molar maf poly- (and oligo-)olefins obtained with MAO-activat


Scheme 1. Potential functionalizations and applications for specialty chemicals on the basis of oligopropene (olefin oligomers in general) (adapted from[7,47]). Further details: reaction with maleic anhydride and reaction of the obtained succinic anhydride with diamine-terminated polyamide-6,6 to give triblockcopolymer dispersing agents, see[47,48,65]; thiol-terminated oligopropenes as chain transfer agents in radical polymerizations, see[47,49]; hydroxy-terminatedoligopropene as initiator for caprolactone polymerization, see[47]; methacrylate-terminated oligopropene, macromonomer for graft copolymers, see[47,49].

zirconocene systems generally decrease with the concen-tration of the MAO cocatalyst. Addition of AlMe3 alsodecreases the molecular weight due to the action of TMA asa chain transfer agent[85,86].

2. Oligomer analyses

A presentation of methods on oligo-olefin analyses is help-ful as the results from these analyses form the basis of manymechanistic conclusions drawn from the oligomer synthesesby metallocene and related catalysts which will subsequentlybe presented in Sections3 and 4.

2.1. Molecular weight determinations

Molecular weight determinations of olefin oligomers aretypically based on either gel permeation/size exclusion chro-matography (GPC/SEC),1H NMR spectroscopy, gas chro-matography (GC), viscosity and/or cryoscopy measurements.

2.1.1. Gel permeation/size exclusion chromatographyGPC/SEC is widely used for determining the molar mass

and distribution of polymer samples. Recent work has, how-ever, shown the use of GPC together with the universal cal-ibration method to be problematic when molar masses aresmall (


dispersityQ = Mw/Mn can, however, only be determined byGPC.

2.1.2. 1H NMR spectroscopyAssuming that each oligomer possesses a double-bond end

group, the ratio of olefinic protons to alkyl protons can bedetermined by integration of the1H NMR spectrum. Fromthis an average chain length is derived which translates intothe number average molar mass,Mn [89]. Furthermore, therelative content of different types of double bonds (if present)can be determined. Among the methods described here, NMRspectroscopy is one of the fastest and simplest procedure,requiring only a small amount of sample.

2.1.3. Gas chromatography, GC and GC/massspectrometry, GC/MS

For oligomeric fractions which are sufficiently volatileand can be evaporated without decomposition at elevatedtemperatures, a GC (with flame ionization detection, FID)or GC/MS analysis is often used to separate the differentoligomers, to determine their individual molecular weight(by MS) and, hence, the degree of oligomerization, and toestablish their identity. The latter is either done by MS anal-ysis or by comparison with separately synthesized or other-wise available reference compounds. GC or GC/MS anal-ysis is often applied to low oligomeric products, that is,o ers,t nsw sep-a tionc es aq rento r-a Co

2-

t ahy-d re-m d totw -l radap low-i[

2t or

v er av-e n-d s seens es oft ibra-t red

(300–500 mg) and the lower solubility of the higher molarmass oligomers proves disadvantageously. Cryoscopy can becarried out in benzene, vapor pressure osmometry in chloro-form (at 25◦C) or in methylcyclohexane (at 50◦C) [96].

2.1.6. Raman spectroscopyFor 1-decene oligomers the relative content of double

bonds could be determined by the integrated intensities of theC C vibrational bands around 1660 cm−1 when referencedto the intensity ofν(C C) of trans-4-decene[97]. Assumingthat each oligomer chain contains a double bond, the rela-tive intensity of the CC vibration can be compared to anexternal standard and is then inversely proportional to theaverage chain length from whichMn can be obtained. Forthe vinylidene-terminated propene and 1-hexene oligomers,2-ethyl-1-butene, H2C C(Et)CH2CH3, was used as a refer-ence whose molar mass corresponds to a dimer or monomerunit, respectively. The CC stretch lies at 1650 cm−1 for thevinylidene (1) and at 1642 cm−1 for the vinyl group (2). Al-ternatively, IR spectroscopy could be applied for the samepurpose by using the peaks at 888 cm−1 (vinylidene) and910, 992 cm−1 (vinyl) [98]. The error in the reproducibilityof the peak integral was found to be 3% for the standard ma-terial and up to 10% for the oligomers, depending on the peakarea[81].

2

2por-

t sign-mo -f witht ciesaN e-a

( dg res-e vinyle ofi tedi

2d

har-a ers[ atem ion-i rgea them DIi hich

ligomers formed at early reaction stages (dimers, trimetramers, etc.). Typically, long (>25 m) capillary columith silicone stationary phases are used for the GCration of the unpolar oligomers. If complete evaporaan be ensured integration of the GC peaks providuantitative measure to the relative content of the diffeligomers[45,54,63,73,75,79,85,90–92]. The number avege oligomerization degreePn can be approximated from Gligomer distributions[77].

.1.4. ViscosityThe intrinsic viscosity (η in dL g−1) of oligomer solu

ions in decalin (decahydronaphthalene) or tetralin (tetrronaphthalene) at 135◦C can be obtained from measuents with a Ubbelhode type viscometer and converte

he viscosity average molecular weight,Mη [93] or to theeight average molecular mass,Mw. The latter is calcu

ated by using, for example, the Mark–Houwing–Sakuarameters for atactic polypropene shown in the fol

ng equation [�] = 1.85× 10−4 Mw0.73dL g−1 [94,100] or�] = 1.10× 10−4 Mw0.80dL g−1 [95].

.1.5. Cryoscopy and vapor pressure osmometryThe molar mass obtained through the melting poin

apor pressure depression corresponds to the numbrage molar mass,Mn. With the help of molecular staards, a considerable deviation to lower mass values watarting above 500 g/mol. Hence, the experimental valuhe oligomers had to be corrected according to this calion [81]. In addition, the high amount of sample requi

.2. End group analyses

.2.1. 1H and 13C NMR spectroscopyThe (start and) end group structure is of uttermost im

ance in elucidating the chain transfer mechanism. Asent of the end group functionality is typically based on1H,ccasionally also on13C NMR. Scheme 2collects the dif

erent possible end groups for oligopropenes togetherheir originating chain transfer reaction or chain start spend gives the NMR chemical shifts.Fig. 1 illustrates the1HMR spectra in the olefinic region of a mainly vinylidennd allyl-terminated oligopropene.

To assess the origin of, for example, aniso-butyl groupeither as a start group from�-Me elimination or as an enroup from chain transfer to Al) one can check for the pnce of and the correlated intensity to the signals of thend group (from�-Me elimination). For the correlation

ntensity in13C NMR it is appropriate to compare correlantensities between spectra of different oligomers.

.2.2. MALDI–TOF MS (matrix assisted laseresorption ionization–time of flight mass spectrometry)

This method has proven quite successful for the ccterization of large biopolymers and synthetic polym

103,104]. It offers the possibility to determine the accurolecular mass and the molar mass distribution. The

zation technique allows for the investigation of very land/or thermolabile molecules without fragmentation ofolecular ions. A prerequisite for the application of MAL

s the presence of functional groups in the polymer w


are necessary for the adduct formation with silver (sodium orpotassium) ions used for the ionization. Hence, polyolefinsare in principle not susceptible to the MALDI technique. Thereason that MALDI–TOF MS also promises a potential foranalysis of olefin oligomers is because of their double-bond

end-group functionality. MALDI–TOF MS investigations ofsuch unpolar samples are quite rare, so far[53,81,82].

Fig. 2 shows a typical MALDI–TOF mass spectrumfor an oligopropene and an oligo-1-hexene material. Thepeak separation corresponds to the monomer mass. From

Sti

cheme 2.1H and13C NMR chemical shifts (relative to TMS;1H olefinic region oogether with their originating chain transfer reaction (P = continuing polymerso-butyl group is sometimes also referred to asiso-propyl.

nly;13C assignments in italics) for common poly-/oligo-propene end groups/oligomer chain). NMR chemical shifts from Ref.[11,79,81,82,99,100–102]. The


Fig. 1. 1H NMR spectra of the olefinic region of a mainlyvinylidene-terminated (top) and vinyl/allyl-terminated (bottom) oligo-propene (200 MHz, CDCl3). The protons of the vinylidene double bond ap-pear as broad singlets at 4.67 and 4.75 ppm. The weak multiplet at 5.45 ppm(*) in the top spectrum is assigned to acis-2-butenyl group arising from�-H elimination after a secondary, 2,1-insertion. The allyl group protonsfeature two overlapping doublets at 4.95–5.05 ppm and a multiplet (ddt) at5.70–5.92 ppm. The vinyl/allyl-terminated oligopropene was obtained withcatalyst (C5Me5)2ZrCl2/MAO [81,82,102].

the mass/charge minus a whole-numbered multiple of themonomer mass, the residual mass is obtained which corre-sponds to the start group. Normally the intensity of masspeaks in mixtures is proportional to the molar amount ofeach species and therefore the mass spectra should yieldMnandMw/Mn (Q). Indeed the mass spectra inFig. 2 representa mass distribution curve. However, it is apparent that themethod and the apparative conditions impose severe limita-tions with respect to interpreting the olefin oligomer spectrain this direction. The vacuum which is applied before ioniza-tion already removes the very low molar mass content fromthe sample, which is rather critical here. At the same time, thehigh molar mass end of the sample is also not well reproducedbecause of the single double-bond functionality together withthe detection of the molecular ion as a metal adduct. Thus,olefin oligomers investigated with MALDI–TOF MS give atoo small dispersionQ which does not reflect the real mo-lar mass distribution. Still, MALDI–TOF is able to providevaluable information concerning the start and end group ofthe oligomers and the mechanism of chain transfer as is il-lustrated inFig. 3together withScheme 3 [81,82].

Evidence for the�-methyl elimination is obtained in theMALDI–TOF mass spectra by anMi − 14 andMi + 14 peakaroundMi in Fig. 3 (Mi = multiples of 42 for propene). The

Fig. 2. MALDI–TOF mass spectra for samples of oligopropene (top) andoligo-1-hexene (bottom). Peaks correspond to the silver adduct of Mi(i = degree of oligomerization, number of monomer units) (a.i.: arbitraryintensity)[81,82].

Mi − 14 peak originates from an oligomer with a [Zr]–Hstart group and chain termination by�-CH3 elimination(Scheme 3b). Since the latter gives rise to Zr–CH3 start groupsan Mi + 14 peak is seen for those oligomers which are thenterminated by�-H elimination (Scheme 3c). The oligomerswhich start with [Zr]–CH3 and end with�-CH3 eliminationfeature theMi peak (just like those with a [Zr]–H start groupand�-H elimination) (Scheme 3a and d).

MALDI–TOF MS can also support the formulation of achain-transfer to aluminum.Fig. 3, bottom displays a samplespectrum of an oligo-1-hexene which featuresMi + 14 peaksin addition toMi peaks. Chain-transfer to aluminum will giverise to saturated end groups which do not show as peaks inthe MALDI–TOF MS due to lack of the double-bond func-tionality (Scheme 3f and h). TheMi andMi + 14 peaks corre-spond to oligomers from [Zr]–H and [Zr]–CH3 start groups,respectively, which are terminated by�-hydrogen elimina-tion (Scheme 3e and g). The relative intensities of theMiandMi + 14 peaks are a direct measure of the probability orpercentage of the chain-termination pathway.

It is, however, difficult to assess the importance ofchain transfer to aluminum with MALDI–TOF MS if�-methyl elimination is operating simultaneously. The ratio ofthe Mi − 14 andMi + 14 peak might be used as a measure.


Fig. 3. MALDI–TOF mass spectra for samples of an oligopropene with�-methyl elimination (top, note the Mi + 14 and Mi − 14 peaks) and an oligo-1-hexene with chain transfer to aluminum (bottom, note the Mi + 14 peak);see alsoScheme 3. Peaks correspond to the silver adduct ofMi (i = degree ofoligomerization, number of monomer units) (a.i.: arbitrary intensity)[81,82].

However, both peaks possess a different double-bond(vinyl/allyl versus vinylidene) which may not have the sameion-coordination and, thus, ionization probability[81,82].

3. Metallocene complexes for olefin, alkyne andsilane oligomerization

3.1. Ethene oligomerization

Wang et al. described the catalytic cyclo-oligomerizationof ethene by (C5H5)2ZrCl2/AlEt3/MgEt2 (ratio 1:100:10;150◦C,pC2H4 = 1.4 MPa) to give 51% of the cyclic oligomersexo-methylenecyclopentane (41%), methylcyclopentane(3%), 1-methyl-1-ethylcyclopentane (4%) and vinylcyclo-hexane (3%) together with open-chain alkenes [Eq.(1)]. With(C5Me5)2ZrCl2, (Ind)2ZrCl2 or (C5H5)2TiCl2 in place of(C5H5)2ZrCl2 the selectivity for exo-methylenecyclopentanedropped to 1–8%[105]. Also ethylalumoxane/AliBu3 couldbe used as a cocatalyst for (C5H5)2ZrCl2 [Eq. (1)].With ethylalumoxane/AlEt3 as cocatalyst under optimalconditions, the oligomerization of ethene gave methylenecy-clopentane and C4–C10 chain olefins in a ratio of 45/55[106].

The oligomerization of ethene by (C5H5)2ZrL2 (L = Cl,Me, OC6H4-p-Me)/ethylalumoxane/AlEt3 (Al:Zr = 100,150◦C, pC2H4 = 1.4 MPa) gave methylenecyclopentanein 39% selectivity which could be further improved to43% under optimal conditions by addition of C5H5N to(C5H5)2ZrCl2/AlEt3 [107,108].

(1)

Sun described a mixed 8-aminoquinoline nickeldichloride/(C5H5)2ZrCl2 catalyst supported on silica toproduce branched polyethylene with oligomer chainsremaining in the final solution. With decreasing amount ofNi catalyst in the supported catalyst, the molecular chainsof oligomers in the resulting solution became shorter, while�-olefin selectivity increased[109].

Carpentier et al. used in situ combinations of(C5Me5)2Nd(�-Cl)2Li(OEt2)2 and Mg(Et)(nBu) orMg(nhexyl)2 to obtain oligoethylenes withM up to2 -a1t

sa rsionM ns[ m-b tedt ieldd hy-d n(MM ayoMSL

500 and narrow dispersionMw/Mn < 1.10 in moderte activity (79× 103 g molNd−1 h−1 atm−1 at 80◦C,atm). Combinations of rac-Me2Si(C5H2-3-SiMe3-4-

Bu)2Nd(�-Cl)2Li(THF)2/Mg(Et)(nBu) or Mg(nhexyl)2howed decreased activity compared to the C5Me5-Ndnalog but the oligoethylenes had also a narrow dispew/Mn < 1.2 and higher contents of vinyl terminatio

110]. The trivalent neodymium complexes when coined in situ with a dialkylmagnesium cocatalyst, initia

he oligomerization of ethene (and 1-octene) to yi(oligoalkyl)magnesium species which were finallyrolyzed to oligomers [Eq.(2)]. Ethylene oligomerizatioMn = 400–5000) was best achieved with the complexesrac-e2Si(C5H2-2,4-{SiMe3}2)2Nd(�-Cl)2Li(THF)2 and (C5e5)2Nd(�-Cl)2Li(OEt2)2 as significant catalyst decccurred with the also investigated complexesrac-e2Si(C5H2-3-SiMe3-4-tBu)2Nd(�-Cl)2Li(THF)2 and Me2i(C5H2-2,4-{SiMe3}2)(C5H2-3,4-{SiMe3}2)Nd(�-Cl)2i(THF)2 [111].

(2)


Gotz et al. obtained small amount of waxes due to ethy-lene oligomerization during high pressure polymerizationof ethylene at 150 MPa, 180◦C and higher Al concentra-tions with (C5H5)2ZrCl2 or Ph2C(C5H4)(Flu′)ZrCl2/Al iBu3(=TIBA)/AlEt 3/[PhNHMe2][B(C6F5)4] (4) [112].

Bravaya et al. described low-molecular weightpolyethylene from Me2Si(C5H4)2ZrMe2/Al iBu3/[Ph3C]-or [PhNHMe2]-[B(C6F5)4] and from Me2Si(C5H4)2ZrCl2/Al iBu3/[PhNHMe2][B(C6F5)4] (5). Analysis of thepolymer products by1H NMR, IR, GPC and kinetics of thepolymerization suggested that the active site contains AliBu3as a heteronuclear ZrAl bridged cationic complex andthat the growing polymer chain is intensely transferred tothe monomer withkp/kt = 350–500 for the above dimethyl-complex catalytic system[113,114].

He et al. studied the ethylene oligomerization activitiesand selectivities of (Ind)2Zr(O-C6H4-p-Me)2/AlEtnCl3−n(n = 1, 1.5, 2) at different reaction and aging temperaturesas well as at various reaction times. The activity increasedwith the ethyl amount (n) in the cocatalyst to a maximumof 1666 g gZr−1 h−1. The selectivity for C4–C10 olefins was92.7%, the selectivity for linear�-C4–C10 olefins was 86%[115].

Siedle et al. carried out a mechanistic study on thehomogeneous catalyst formed from (C5H5)2ZrMe2/MAO intoluene. They relied on ethene oligomerization to investigatethe chain termination which predominantly involved transferto Al in exchange with the Al–Me group together with a con-comitant chain transfer by�-bond metathesis of the Zr-boundgrowing chain with ethene. From the former, the terminalAl-chain bonds of the Al-bound oligomer are finally cleavedon hydrolytic workup to produce saturated end groups. Thelatter �-bond metathesis yields the new initiating species[(C5H5)2Zr-CH CH2]+ from which mono-unsaturated Cn

Ss

cheme 3. Correlation between start and end group for the different chain teen in MALDI–TOF MS. Saturated oligomers do not give an MS peak with

ransfer reactions. Only oligomers with vinylidine or vinyl/allyl end groups can bethis method[81,82].


(n = even) alkenes are obtained. Chain transfer by�-Helimination was detectable but relatively insignificantunder the conditions employed. For propene and 1-hexene�-H elimination was the predominant chain transfer step,however. Chain transfer was fast relative to propagationso that the products were low molecular weight oligomers[80].

Conti et al. described the formation of etheneoligomers (Mn = 4000 g mol−1, Mw/Mn = 4.25, sample con-stituted of solid PE and oligomers of C6–C30 dis-tribution) with (C5HPh4)2ZrCl2/MAO in lower activity(75× 103 g molZr−1 h−1) than in the case of unsubstituted ormethyl-substituted ligands (186 or 372× 103 g molZr−1 h−1,respectively). Steric and electronic effects by the phenylsubstituents of this “bulky” cyclopentadienyl ligand[116] were seen responsible for this catalytic behavior[117].

Fink and Schnell obtained polymerization rates and de-veloped mechanistic suggestions for Ti alkylation, activationequilibria, Ti-olefin� complex formation, chain growth attitanium and Ti→ Al chain transfer processes based on ki-netic studies and the analysis of ethene oligomer distribu-tions from short reaction times with (C5H5)2Ti(R)Cl/AlEtCl2(R = Et,nPr,nBu,n-pentyl,n-hexyl) [84]. Reichert and Meyerhad carried out similar short time ethene polymerizationswith (C H ) Ti(Et)Cl/AlEtCl to arrive at ethene oligomersw rtions

3

AO( o-t andw enep( olyc talyst( fe5

gedz( llyc ea pa-r haint enem unit[

3

lyst( ste ad-

ing to exclusively vinylidene-terminated oligopropenes[43,47–49,78,79,119–121]. Actually all members ofthe methyl-substituted series (C5H5−nMen)2ZrCl2up to (C5Me5)2ZrCl2 as well as the mixed methyl-substituted zirconocenes (C5H5)(C5HMe4)ZrCl2 and(C5H5)(C5Me5)ZrCl2 give atactic propene oligomers incombination with MAO[41,79,102].

Mülhaupt et al. showed that addition of the Lewis-acids trimethyl- or triphenylboroxine, (BOMe)3 or(BOPh)3, respectively, or phenylboronic acid, PhB(OH)2to (C5H5)2ZrCl2/MAO could substantially enhancecatalyst activities and affect molecular weight. Froman activity of 3.68× 105 g molZr−1 atm−1 h−1 an in-crease to 11.12× 105 with (BOMe)3, 5.56× 105 with(BOPh)3 and 4.95× 105 g molZr−1 atm−1 h−1 withPhB(OH)2 was found under otherwise identical con-ditions (temperature = 40◦C, propene pressure 2 bar,[Zr] = 50× 10−6 mol l−1, Al:Zr = 500) and with a molarBmodifier:Zr ratio of 25. In the same order the molecularweight Mn increased from 300 g mol−1 without modifier to570, 830 and 420 g mol−1 [119].

Jacobs et al. described the immobilization of(C5H5)2ZrMe2 on zeolite MCM-41 whose silanolgroups were pre-reacted with B(C6F5)3/PhNMe2 toresult in the heterogeneous catalyst [{MCM-41-O}-B(C F ) ]−[(C H ) ZrMe]+. Activities of this heteroge-n le tohB -a

ins am ctiveh ew neo

enebr car-b ,3,4-t ivity6p era-t

5 5 2 2hose analyses yielded rate constants for the first inseteps of the monomer[83].

.2. Ethene/α-olefin co-oligomerization

Van Looveren et al. used the in situ synthesis of Mfrom AlMe3) on the crystalline structure of phosphungstate heteropoly acids to generate a highly activeeakly coordinating compound for the ansa-metallocre-catalyst C2H4(Ind′)2ZrCl2, Me2Si(Ind′)2ZrMe2 orC5H5)2ZrMe2. The combination of the alumoxo-heteropompound and the metallocene gave a highly active ca∼2–4× 106 g molZr−1 h−1) for the co-oligomerization othene and propene (Mn = 660–1300 g mol−1, Mw/Mn ≈ 2,3–58 mol.% propene)[67,118].

Ciardelli et al. tested the capacity of the nonbridirconocene complexes (C5H5)2ZrMe2, (Ind)2ZrCl2 andC5Me5)2ZrCl2 activated with MAO to produce structuraontrolled oligomers from ethene/�-olefin mixtures (propennd higher olefins) by varying the different reactionameters (temperature, pressure, Al:Zr ratio). The cransfer reaction is most probably induced by the proponomer when the propagating chain end is a propene

56].

.3. Propene oligomerization

The achiral prototypical metallocene cataC5H5)2ZrCl2/MAO appears to be one of the momployed systems for propene oligomerization le

6 5 3 5 5 2eous propene oligomerization catalyst were comparabomogeneous (C5H5)2ZrCl2/{AlMe3, MAO, B(C6F5)4− or(C6F5)3} systems (∼105 g molZr−1 h−1) and over 90% 1lkenes with Schulz-Flory carbon number distribution[122].

Van Looveren et al. similarly anchored MAO by theitu hydrolysis of AlMe3 on the internal pore walls ofesoporous MCM-41 support to generate a highly aost for (C5H5)2ZrMe2 in the oligomerization of propenith a typical Schulz-Flory distribution of the propeligomers[57,58].

Wang et al. catalyzed the oligomerization of propyly (C5H5)2ZrCl2/ethylalumoxane or AlEt3 at low Al/Zratios and high temperature to afford the trimericocyclic products 1,1,3-trimethylcyclopentane and 1,1

etramethylcyclopentane (combined optimized select1%), along with C6 and C9 chain olefins [Eq.(3)]. Theroduct distribution was affected by the reaction temp

ure, propylene pressure, Al/Zr ratio and reaction time[123].

(3)


Erker et al. describedansa-metallocenes with long(CH2)n chains, namely (CH2)n(C5H4)2ZrCl2 (6)/MAO

with n = 9, 12 for the oligomerization of propene with ac-tivities of 4.56 and 3.32× 105 g molZr−1 h−1 bar−1 [124].

The very rigid compounds (4-cyclopentdienylidene-7-R-4,7-dimethyl-4,5,6,7-tetrahydroindenyl)MX2 (R = Me,MX2 = Zr- or HfMe2; R = Ph, nBu, allyl, -(CH2)3-9-BBN,MX2 = ZrCl2) (7) exhibit a large opening angle in front ofthe metal (gap aperture) and produce low molecular weightpropene oligomers (Mη = 830–1900 g mol−1) when activatedwith MAO between −20 and 0◦C; activities between0.15–11× 103 g gZr−1 h−1 (Al:Zr = 550–860)[50,125].

The chiral, meso-like ansa-zirconocene complex8/MAOproduces oligopropenes (Mη = 4400 – 380 g mol−1) between0 and 80◦C (pC3H6 = 2–20 bar, Al:Zr = 1100–2000) with ac-tivities between 0.34 and 1.2× 103 g gZr−1 h−1 bar−1, re-spectively[44].

Schaverien et al. reported on the class of ethylene-bis(2-indenyl)zirconocenes, (C2H4)(2-Ind′)2ZrCl2 with var-ious methyl substitution patterns in the 1- and 3-position andphenyl-substitution in the 4-position (Scheme 4). When acti-vated with MAO (Al:Zr between 1000 and 40,000) these com-pounds oligomerized propene (Mn from 1H NMR between400 and 5500 g mol−1) with activities between∼2× 103 and6.0× 104 g gZr−1 h−1. Addition of 2–3 mol% H2 drasticallyincreased the activities by a factor of >100 for the same cata-lyst and up to 6.2× 106 g gZr−1 h−1. The formation of oligo-p on of

Scheme 4. End group analysis of oligopropenes by NMR base

ropenes was exploited for a comprehensive identificati

d on (mostly ethylene) bridged bis(2-indenyl)zirconocenes[126].


the end groups, thereby providing information on the chaintransfer mechanism. In the absence of H2 n-propyl (start)and vinylidene end groups characteristic of 1,2-insertion intoZr H and�-H elimination to the metal or�-H transfer tomonomer are found. Strong signals were also observed forisobutyl end groups, indicating chain transfer to Al (andeventual hydrolysis) as a major termination pathway, com-mensurate with the high Al:Zr ratio and the low activity(isobutyl start groups then also arise from obtaining ZrMestart species by this transfer route). A significant percentageof 2,1-regioerrors with some of the complexes gaven-butylend groups after transfer to Al. Furthermore small amountsof allyl end groups (from�-Me elimination) were observedthroughout. Relatively high levels ofcis- or trans-2-butenyland 4-butenyl end groups were also formed by�-H transferto monomer after a 2,1-insertion. Addition of H2 reduced the2,1-insertions and increased the saturated propene oligomersdue to chain transfer to H2 [126].

Bravaya et al. found that an increase in the Al:Zr ratio inthe catalytic system (2-PhInd′)2ZrMe2 (9)/Al iBu3 from 50to 300 led to an enhancement of the molecular weight of thepolypropene samples from oligomeric products to a viscosityaverage molecular weight of 220,000 g mol−1 [127].

mp-l(( rr lids ilyp -i

edf dc rs(a rbcc eric

polypropene[131].

Horton et al. employed phospholyl, C4PH2-2,5-R2containing metallocene complexes of the type (C4PH2-2,5-R2)2ZrCl2 and (C5H5)(C4PH2-2,5-R2)ZrCl2 (R = H,cyclo-propyl,tBu, SiMe3, Ph) (12) in propene polymerizationusing MAO as cocatalyst. Atactic polypropene withMn vary-ing from 450 to >20,000 g mol−1 and vinylidene end groupswas obtained with activities up to 1.7× 105 g gzr−1 h−1.The degree of polymerization increased steadily in theseries dialkyl < alkyl, phenyl < diphenyl suggesting a higherimportance of electronic over steric factors forMn [132].

Janiak et al. compared the activities for zirconoceneswith the tetramethyl-substituted phospholyl ligand C4PMe4to the activities of zirconocenes with the tetramethyl- andpentamethyl-cyclopentadienyl ligand C5HMe4 and C5Me5,r lcwt nes

S pholylcomplexes for oligomerization activity and chain transfer reactions.

Mu et al. noted that the unbridged zirconocene coexes (C5H3-1,2-Ph2)2ZrCl2, (C5H2-1,2-Ph2-4-Me)2ZrCl2,C5H2-1,2-Ph2-4-Pr)2ZrCl2, (C5H2-1,2,4-Ph2)2ZrCl2, andC5H2-1,2-Ph2-4-cyclohexyl)2ZrCl2 (10), despite theiacemic-like, near C2 symmetrical conformation in the sotate, when activated with MAO gave only atactic, oropene oligomers (Mn = 1000 g mol−1), albeit at high activ

ties (1.8–4.2× 106 g molzr−1 h−1, Al:Zr = 1500)[128–130].

Similarly, Erker et al. found that the catalyst derivrom the 2-(5′-methyl-2′-N-methylpyrrolyl)-substituteomplex (Ind′)2ZrCl2 (11) yielded only propene oligomeMw = 3600 g mol−1, Al:Zr = 991, temperature −20◦C,ctivity 5.9× 104 g molzr−1 bar−1 h−1) although similais(2-arylindenyl)- and bis(2-hetarylindenyl)ZrCl2/MAOatalysts showed a C2-symmetric rac-like metalloceneonformation in the solid state and gave elastom

espectively (Scheme 5). Activities for the phospholyomplexes (C5H5)(C4PMe4)ZrCl2 and (C4PMe4)2ZrCl2ere much lower (0.1× 105 g molzr−1 h−1), by a fac-

or of ∼100 than those of the per-methyl zirconoce

cheme 5. Comparison of per-methyl zirconocenes with related phos


(6–13× 105 g molzr−1 h−1, 50◦C, c0(C3H6) = 1.75 mol l−1,[Zr] = 3 × 10−5 mol× l−1 with Al:Zr = 1000) [81,82,102].This is explained by formation of a P–Al adduct[132,133]with the bulky “Al-substituent” becoming situated close tothe maximum opening angle/gap aperture of the zirconium,that is, close to the meridional centroid–Zr–centroid angle(seeScheme 5). Such a bulky substituent in this position thenblocks the pathway for the insertion of propene into the ZrCbond (as is known foransa-metallocenes with C5-ring sub-stituents in this position)[82,102,134]. Molecular weightsfor the oligopropenes obtained with the phospholyl systemswere betweenMn = 300–380 g mol−1 and at the low end ofthe permethyl zirconocene series (Mn = 320–1500 g mol−1)with only decamethylzirconocene, (C5Me5)2ZrCl2 yieldinga comparable low molar mass oligomer under the same con-ditions. With the phospholyl systems and with the symmetri-cal per-methyl zirconocenes (C5H5−nMen)2ZrCl2 (n = 4, 5)both vinylidene and vinyl/allyl-terminated oligopropenes arefound. The vinyl:vinylidene ratio depends on the ligand ormetal complex and increases from (C5H5)(C4PMe4)ZrCl2(21:79) and (C5HMe4)2ZrCl2 (35:65) over (C4PMe4)2ZrCl2(39:61) to (C5Me5)2ZrCl2 (91:9). Furthermore, compari-son of the molecular weight determinations by1H NMRand GPC together with MALDI–TOF MS shows that thephospholyl complexes exhibit some chain transfer to Al[

( f3a tivityww(

nedC f5[

m rsc trod sto imed tiona ows[

Pino et al. polymerized propene in the presence of hy-drogen with asymmetric (R)-C2H4(Ind′H4)2ZrMe2/MAO toarrive at hydrooligomers (14) together with solid high poly-mer. The oligomers were separated and fractionated. Therotation of all the fractions was positive. Specific isomersof the trimers 2,4-dimethylheptane and 4-methyloctane andthe tetramers 2,4,6-trimethylnonane and 4,6-dimethyldecanecould be identified. Then-butyl groups indicate the chain ter-mination by hydrogenolysis after a 2,1-insertion[96].

In a mechanistic study Siedle et al. used(C5H5)2ZrMe2/MAO and propene oligomerization intoluene to find�-H elimination as the predominant chaintransfer which was different than for ethene oligomerizationwith the same system (see above)[80].

e

te,

yl

is

lo-

81,82,102].Ciardelli et al. tested (C5H5)2ZrMe2, (Ind)2ZrCl2 and

C5Me5)2ZrCl2 activated with MAO at Al:Zr ratios o00–3000, at two different temperatures (20 and 70◦C)nd 1 or 5 bar propene pressure. The highest acas found for (C5H5)2ZrMe2 at Al:Zr = 3000 at 70◦Cith 2.35× 106 g molzr−1 h−1, oligomer range C6–C30+. For

Ind)2ZrCl2 the oligomers were mostly >C30 [56].Kaminsky et al. obtained over 99% prope

imer in the oligomer mixture (by GC) withrac-2H4(Ind′H4)2ZrCl2/MAO, at a monomer rate o0 mmol h−1, [Zr] = 5 × 10−4 mol l−1, Al:Zr = 200, 60◦C

Eq. (4)] [85].

(4)

With asymmetric (S)-C2H4(Ind′H4)2Zr(O-acetyl-(R)-andelate)2/MAO (13) optically active propene oligome

ould be obtained through enantiomorphic site conuring the C C bond formation. Optical activity is firbserved in the trimers and in higher oligomers, the does not contain a chiral carbon atom. Molar rotangles for the trimer and tetramer mixture were as foll

45,135]:

l

r

Jordan et al. noted that the complex (C5H4Me)2ZrMe(�1-CB11H12) oligomerizes propene under mild conditions. Thcarborane ligand CB11H12− coordinates to Zr via a singleB–H–Zr bridge and is labile so that a free coordination sinecessary for catalytic activity, can be created[136].

Stockland and Jordan found that the reaction of vinchloride withrac-C2H4(Ind′)2ZrMe2/MAO (Al:Zr = 17,000)proceeds by generation ofrac-C2H4(Ind′)2ZrMe+, vinylchloride insertion, and�-Cl elimination to yield propeneand arac-C2H4(Ind′)2ZrCl+-species [(Eq.(5)]. The latteris realkylated by MAO (or its AlMe3 content). This pro-cess is stoichiometric in Al–Me groups. The propeneoligomerized byrac-C2H4(Ind′)2ZrMe+ to atactic oligo-propene (Mn = 500 g mol−1) as a colorless oil with thepredominant chain transfer mechanism being vinyl chride insertion/�-Cl elimination. Similar results were said tohave been obtained withrac-C2H4(Ind′)2ZrCl2/MAO and(C5Me5)2ZrCl2/MAO. In the presence of trace levels of O2,rac-C2H4(Ind′)2ZrMe2/MAO polymerizes vinyl chloride to


poly(vinyl chloride) by a radical mechanism[137,138].

(5)

Janiak et al. used a series oftert-butyl substituted zir-conocene (15)/MAO complexes to illustrate the potentialfor tailoring propene oligomers with extremely narrow mo-lar mass distribution down toMw/Mn ≈ 1.2. This small dis-persion is explained on the basis of a chain-length depen-dent propagation rate. At 50◦C, c0(C3H6) = 1.75 mol l−1and [Zr] = 3× 10−5 mol l−1 with Al:Zr = 1000 activitieswere 4.0× 105 g molzr−1 h−1 for the zirconocenes withthe mono-substituted ligands (C5H5)(C5H4tBu)ZrCl2 and(C5H4tBu)2ZrCl2. A maximum was observed for the mixedzirconocene with one di-substituted ligand, (C5H5)(C5H3-1,3-tBu2)ZrCl2 with 6.7× 105 g molzr−1 h−1. Upon fur-ther increase in steric demand the activities decreased to0.8× 105 g molzr−1 h−1 for (C5H5) (C5H2-1,2,4-tBu3)ZrCl2and 0.01× 105 g molzr−1 h−1 for (C5H3-1,3-tBu2)2ZrCl2.Molecular weights were betweenMn = 400–1300 g mol−1.S nd[

con-d hingzd( inm qui-l atedm mer.A ble tor

M

attap ing

Scheme 6. Differences in monomer enchainments and possible end groupstructures after 1,2- or 2,1-insertion.

13C NMR characterizations of polymers with GC/MS analy-ses of oligomers which resulted from fast H2-induced chaintransfer (hydrooligomerization). This approach allowed forthe evaluation of the ratios between kinetic constants gov-erning the formation of the various possible constitutionaland configurational sequences. Specifically, the ratio for therate of chain propagation via regular 1,2-, primary insertion(kpp) and the rate of regioirregular 2,1-, secondary inser-tion (kps) (Scheme 6), i.e. kps/kpp for the two catalyst sys-temsrac-C2H4(Ind′)2ZrCl2 (kps/kpp = 1.0× 10−2) and rac-Me2Si(Ind′)2ZrCl2/MAO (kps/kpp = 0.8× 10−2) were basedon the ratio between 1,2-, primary-terminated and 2,1-,secondary-terminated chains in the hydrooligomer, measuredby GC[140–143].

Mülhaupt et al. used propene oligomerization studies with(C5H5)2ZrCl2/MAO to demonstrate that reversible deactiva-tions which are second order relative to the zirconium activesite concentration account for the decay of the polymerizationrate. With MAO injection during the reaction the polymer-ization rate can be enhanced again. A sequence of dynamicequilibria involving the formation of active cationic metal-locene intermediates as well as inactive zirconocene speciesis proposed[119,120].

Mülhaupt et al. also found that propene polymeriza-tion with rac-Me2Si(benz[e]indenyl)2ZrCl2 and with rac-Mr t of2p mei siso ro-d cted.T havea ump-tZ ser-t st-i n. Int ,1-p

olely vinylidene terminated oligomers were fou43,81,82,102].

Resconi et al. observed that under “starved catalyst”itions, that is, with monomer concentration approacero ([C3H6] → 0 mol l−1) rac-C2H4(Ind′)2ZrCl2/MAO pro-uces atactic propene oligomers (Mn = 1080 g mol−1) [Eq.6)]. The loss of stereospecificity with the decreaseonomer concentration is explained by postulating an e

ibrium between an isospecific site having a coordinonomer and an aspecific one without coordinated monos a consequence the aspecific site is assumed to be a

acemize the chiral carbon of the last inserted unit[139].

(6)

echanistic studies

Busico et al. carried out kinetic studies of Ziegler–Nolyinsertions with metallocene/MAO catalysts combin

e2Si(2-Me-benz[e]indenyl)2ZrCl2 in the presence of H2educed the polymer molecular weight and the amoun,1-units in the polymer [Eq.(7)]. The NMR signals ofn-ropyl (start) andn-butyl (end) groups had almost the sa

ntensity. Then-butyl groups result from the hydrogenolyf Zr-2,1-units. Hydrogenolysis of a Zr-1,2-unit would puce isopropyl end groups which could hardly be detehis shows that most metallocenes with dormant sites2,1-inserted propene as the last unit (under the ass

ion of a similar rate for the chain transfer to H2 from ar-2,1- and a Zr-1,2-unit). The subsequent monomer in

ion at a sterically hindered “dormant” site with a 2,1-lanserted unit is slower than the hydrogenolysis reactiohe absence of H2 such dormant sites, resulting from 2ropene insertion, can be reactivated either by�-H transfer to


zirconium or to propene, both yielding 2-butenyl end groups,or by 1,2-insertion of propene, yielding head-to-head regio-irregularities[144].

(7)

Mise et al. observed the formation of “abnor-mal” oligomers in the oligomerization of propene by(C5Me5)2MCl2/MAO catalysts (M = Zr, Hf) with the prod-uct mixture analyzed by quantitative GC/MS and1H NMRa witha r1 ad tof ag ofp ne( tiveo

elimination. Oligomers with a vinylidene end group, such asthe dimer 2,4-dimethyl-1-pentene (e) and the trimer 2,4,6-trimethyl-1-heptene (f) start with an MMe bond and ter-minate with the usual�-H elimination. Finally, an M Hstart group and�-H transfer termination was the case in thedimer 2-methyl-1-pentene (g) and the trimer 2,4-dimethyl-1-heptene (h) (Scheme 7) [79].

Resconi et al. used propene oligomerization (at 50◦C)in a mechanistic study on chain transfer via�-CH3 elim-ination by (C5Me5)2MCl2/MAO catalysts (M = Zr, Hf).GC–MS,1H and13C NMR analyses of the oligomers showedthe oligomeric products to be mainly allyl- and isobutyl-terminated (1/1 ratio), the former being the end, the latterthe start group, respectively. The terminal group structuresare produced by first monomer insertion into the MCH3bond and then chain transfer by�-CH3 elimination (c, d inScheme 7). Only small amounts of vinylidene end groupsfrom �-H elimination were found. The allyl:vinylidene ratiowas 92:8 for Zr and 98:2 for Hf[77].

Teuben et al. based a mechanistic study on the oligomer-ization of propene by [C5Me5)2MMe(THT)]+[BPh4]−(M = Zr, Hf, THT = tetrahydrothiophene) in N,N-dimethylaniline. At room temperature for M = Zr a ratherbroad molecular weight distribution is obtained (C6–C24),whereas for M = Hf only one dimer (4-methyl-1-pentene)and one trimer (4,6-dimethyl-1-heptene) are formed. Witha sitions aref -p ightd uresb heS entsf .

S nd the ep

nd separated in part by vacuum distillation. Oligomersvinyl/allyl end group (seeScheme 7), such as the dime

-pentene (a) and the trimer 4-methyl-1-heptene (b) horm from �-Me transfer termination (elimination) fromrowing carbon chain which was initiated by insertionropene into an MH bond. The dimer 4-methyl-1-pentec) and the trimer 4,6-dimethyl-1-heptene (d) are indicaf insertion into an M Me bond with termination from�-Me

cheme 7. Differences in the structure of isolated dimers and trimers aolymerization with (C5Me5)2ZrCl2/MAO [77,79].

n increase in reaction temperature the product compohifts to lower molecular weights. The oligomersormed by �-CH3 elimination from the growing oligoropene chain to the metal center. The molecular weistributions of the oligomers produced at temperatetween 5 and 45◦C are satisfactorily described by tchulz–Flory theory. From this the ratios of rate coeffici

or propagation,kp and termination,kt could be calculated

ir mechanistic interpretation due to�-hydrogen or�-methyl elimination in propen


Values for (�Gp‡–�Gt‡)298 K were calculated as−1.9(3)and −1.4(4) kcal mol−1 for M = Zr and Hf, respectively[75,76].

3.4. 1-Butene oligomerization

Christoffers and Bergman found that (C5H5)2ZrCl2 whenactivated by only 1 equiv. of MAO selectively and catalyti-cally dimerizes 1-butene (and other�-olefins) to 2-ethyl-1-hexene without formation of any higher oligomeric species[Eq. (8)] [51].

(8)

Plattner, Chen et al. prepared a (C5H5)2ZrMe+-cation froma stable solution of the [B(C6F5)4]− salt by electrospray ion-ization tandem mass spectrometry (ESI–MS/MS). The cationwas collided with 1-butene to observe multiple olefin ad-ditions to a gas phase metallocenium cation and to obtaina second-order rate constant ofk ≈ 108–199 l mol−1 s−1 at70◦C [145].

Kaminsky et al. described the product distribu-tion from dimer to heptamers in the oligomeriza-tion of 1-butene as a function of monomer flow, zir-c ThezaCw to0 mi-t .3 to1 o-l ow( tion)a . Fur-t ericp m 30t rr6 (byG

thee ics yi the

dimer does not contain a chiral carbon atom. Molar rotationangles for the trimer and tetramer mixture were as follows[45,85]:

Busico et al. also extended the GC/MS characteriza-tion of saturated propene oligomers (see above[140]) tothe case of 1-butene for the two catalyst systemsrac-C2H4(Ind′)2ZrCl2 and rac-Me2Si(Ind′)2ZrCl2/MAO. Each1-butenen-hydrooligomer contained all diastereomers gen-erated by chain transfer to H2 after a sequence ofn 1,2-monomer insertions, or of (n − 1) 1,2-insertions followed by afinal 2,1-insertion, respectively [Eq.(10)]. The undetectabil-ity of internal 2,1-units confirms the above observations byMülhaupt and co-workers[144]that the subsequent monomeri ,1-l eac-t atedo dbc ryl erea edt a-t d/orH(

onocene concentration and reaction temperature.irconocenes employed wererac-C2H4(Ind′H4)2ZrCl2,symmetric (S)-C2H4(Ind′H4)2Zr(S-binaphtholate), and (S)-2H4(Ind′H4)2Zr(O-acetyl-(R)-mandelate)2 (13) activatedith MAO. With the decrease of flow rate (from 217.8 ml min−1), the shorter oligomers are preferred. Conco

antly, an increase in zirconocene concentration (from 716× 10−5 mol l−1) with simultaneous variation of the m

ar Al:Zr ratio (from 900 to 57) at constant monomer flleading to an effective decrease in monomer concentralso increases the percentage of the shorter oligomers

hermore, the average molecular weight of the oligomroducts decreases with the increasing temperature (fro

o 70◦C). With rac-C2H4(Ind′H4)2ZrCl2/MAO, a monomeate of 50 mmol h−1, [Zr] = 5 × 10−4 mol l−1, Al:Zr = 200 at0◦C an amount of 95% 1-butene dimer was obtainedC) [Eq.(9)] [85].

(9)

Optically active oligomers could be obtained usingnantiomeric zirconocenes, e.g.13, through enantiomorphite control during the CC bond formation. Optical activit

s first observed in the trimers and in higher oligomers,

nsertion at a sterically hindered “dormant” site with a 2ast-inserted unit is slower than the hydrogenolysis rion. Detected at low amounts were terminally unsaturligomers formed by a sequence ofn 1,2-insertions followey H-transfer to the monomer [Eq.(10)]. At low monomeroncentration (


3.5. 1-Pentene oligomerization

Wahner et al. studied the oligomerization of 1-penteneusing (C5H5)2ZrCl2, (C5H5)2HfCl2, (C5Me5)2ZrCl2, rac-C2H4(Ind′)2ZrCl2, Me2Si(Ind′)2ZrCl2, Me2Si(2-Me-benz[e]indenyl)2ZrCl2 [cf. Eq. (7)], (C5H5)2ZrCl{O(Me)CW-(CO)5}, and (C5H5)2ZrCl(OMe), all activated with MAO.Highly depending on the metallocene catalyst, oligomersranging from the dimer of 1-pentene to polymerswith Mw = 149,000 g mol−1 were formed. The system(C5H5)2ZrCl{O(Me)CW(CO)5}/MAO is a highly active cat-alyst for the oligomerization of 1-pentene to low-molecularweight products[146]. (C5H5)2ZrCl2/MAO (Al:Zr = 1000)gave low molar mass oligomers from which the dimer (25%),the trimer (18%) and the tetramer (14%) could be isolated af-ter distillation [Eq.(11)]. The analogous (C5H5)2HfCl2 com-plex yielded an atactic oligomer in the form of a viscousoil (at 60◦C, Mn = 2500 g mol−1). The sterically hinderedcatalyst (C5Me5)2ZrCl2/MAO also gave a viscous oil (at60◦C with Mn = 1000 g mol−1). The catalysts (C5H5)2HfCl2and (C5Me5)2ZrCl2 lead almost exclusively to the forma-tion of vinylidene end groups, whereas the bridged catalystsrac-C2H4(Ind′)2ZrCl2 and Me2Si(Ind′)2ZrCl2 give also in-ternal double bonds of the type RCH=CHR′. These func-tional groups are formed if the chain termination is causedby �-H transfer after a 2,1-“mis”insertion (cf.Scheme 2).A o-1 atrixa tion[

a yti-c -h cies[

o

to evaluate the stereoselectivity for the optically pureC2-symmetric catalysts (R,S)-(BnBp)Y-n-octadeutero-pentyl and (S,R)-(BnBp)Y-n-octadeutero-pentyl (17) andracemic (±)-(BnBp)Y-n-octadeutero-pentyl. Stirring theneat olefin (R)-octadeutero-1-pentene containing (BnBp)Y-CH(SiMe3)2 (18) as catalyst precursor under D2 produced amixture of deutero-oligomers from which the four possibledeuterodimers were isolable by preparative GC [Eq.(13)].These deuterodimers indicate the sense and magnitude ofolefin facial selectivity for insertions into the metal-n-pentylbond with the chiral 1-pentene providing the probe of theabsolute stereochemistry for additions to�-olefin doublebonds[147].

(C -t nan-t -t in ac

3

1Zs

w at-a -l outfi of

MALDI–TOF mass spectrum of low-molar-mass olig-pentene could be recorded using dithranol as mnd adding silver trifluoroacetate to promote ion forma

53].

(11)

In work by Christoffers and Bergman, (C5H5)2ZrCl2 whenctivated by only 1 equiv. of MAO selectively and catalally dimerizes 1-pentene (and other�-olefins) to 2-propyl-1eptene without formation of any higher oligomeric spe

Eq. (12)] [51].

(12)

Gilchrist and Bercaw used optically active (R)-ctadeutero-1-pentene, CD3-CD2-C*HD-CH CD2 (16)

(13)

Kaminsky oligomerized 1-pentene with asymmetricS)-2H4(Ind′H4)2ZrMe2/MAO to obtain optically active pen

ene trimers, tetramers and other oligomers through eiomorphic site control during the CC bond formation. Opical activity starts in the trimer, the dimer does not contahiral carbon atom[85].

.6. 1-Hexene oligomerization

Suzuki et al. reported 1-hexene oligomers (Mn ≈500 g mol−1) with (C5H4tBu)2ZrCl2 or Me2Si(C5H4)2rCl2/MAO in activities of 4 or 8× 105 g molZr−1 h−1, re-pectively[148].

According to Christoffers and Bergman, (C5H5)2ZrCl2hen activated by only 1 equiv. of MAO selectively and clytically dimerizes 1-hexene (and other�-olefins, see be

ow) to 2-butyl-1-octene (=5-methyleneundecane) withormation of any higher oligomeric species [Eq.(14)]. Mon-toring the reaction by1H NMR gave a turnover number


∼6 min−1 after an induction period of∼60 min. It is sug-gested that a Zr-hydrido complex is the actual active species[Eq. (14)]. (C5H5)2Zr(H)Cl/MAO (1:1) also selectivelydimerizes 1-hexene but with a significantly shorter induc-tion period (∼5 min) and a turnover number of∼30 min−1[51].

(14)

Kretschmer et al. applied the dimeric yttrium complex{(2,4,7-Me3-Ind′′′)2Y(�-H)}2 as a catalyst (without any co-activator) in the regioselective dimerization of 1-hexene to5-methyleneundecane in 80% yield (plus 17% of the trimer)at 80◦C [Eq.(15)] [149].

-c lf owm rw ld pen-d[ nta ene(f di-s2 e-mf(f ers 90%y

yl-s df thet andC on

see above,Scheme 5). Activities for the phospholyl com-plexes (C5H5)(C4PMe4)ZrCl2 and (C4PMe4)2ZrCl2were much lower (∼0.1× 106 g molzr−1 h−1), bya factor of ∼50 than those of the permethyl zir-conocenes (3–6.3× 106 g molzr−1 h−1, 50◦C, c0(1-hexene) = 6.9 mol l−1, [Zr] = 3 × 10−6 mol l−1 with Al:Zr =4000) [81,82,102]. This is explained by formation of aP–Al adduct [132,133] with the bulky “Al-substituent”becoming situated close to the maximum opening an-gle/gap aperture of the zirconium, that is, close to themeridional centroid–Zr–centroid angle (seeScheme 5).Such a bulky substituent in this position then blocks thepathway for the insertion of hexene into the ZrC bond (asis known for ansa-metallocenes with C5-ring substituentsin this position) [82,102,134]. Molecular weights for theoligohexenes obtained with the phospholyl systems wereMn = 500 g mol−1 and significantly lower than those of thepermethyl zirconocene series (Mn = 1100–3300 g mol−1).As unsaturated end groups only vinylidene double bonds arefound. Comparison of the molecular weight determinationsby 1H NMR and GPC together with MALDI–TOF MSshows that most of the permethylated complexes exhibitconsiderable chain transfer to Al. In 1-hexene oligomeriza-tion the bis-phospholyl (C4PMe4)2ZrCl2 exhibits about 80%chain transfer to Al. Catalysts based on (C5HMe4)2ZrCl2and (C PMe ) ZrCl give over 60% saturated oligomers.Ts tetra-a( and2 sesw to nsferc way[

sed( int int tionw

f 1-h el ionc ver,t ationo merc

eb tedb -c n-dr hwa ism)i

(15)

Janiak et al. used a series oftert-butyl substituted zironocene/MAO complexes (see15) to illustrate the potentiaor tailoring 1-hexene oligomers with extremely narrolar mass distribution down toMw/Mn ≈ 1.2. Moleculaeights were betweenMn = 200–740 g mol−1. The smalispersion is explained on the basis of a chain-length deent propagation rate. At 50◦C, c0(1-hexene) = 6.9 mol l−1,

Zr] = 3 × 10−6 mol l−1 with Al:Zr = 4000 and a reactioime of 1 h[81,82,102](reaction time was 24 h in[43]) thectivity was highest for the mono-substituted zirconocC5H5)(C5H4tBu)ZrCl2 with 7.9× 106 g molzr−1 h−1ollowed by the mixed zirconocene with the oneubstituted ligand, (C5H5)(C5H3-1,3-tBu2)ZrCl2 with.7× 106 g molzr−1 h−1. Upon further increase in steric dand the activities decrease to 1.7× 106 g molzr−1 h−1

or (C5H4tBu)2ZrCl2, 0.26× 106 g molzr−1 h−1 forC5H5)(C5H2-1,2,4-tBu3)ZrCl2 and 0.3× 106 g molzr−1 h−1or (C5H3-1,3-tBu2)2ZrCl2. The latter zirconocene rathelectively gave the dimer 5-methyleneundecane in 80–ield [43,81,82,102].

The activities for zirconocenes with the tetramethubstituted phospholyl ligand C4PMe4 were also compareor 1-hexene to the activities of zirconocenes withetramethyl- and pentamethyl-cyclopentadienyl lig5HMe4 and C5Me5, respectively (propene comparis

4 4 2 2he decamethylzirconocene (C5Me5)2ZrCl2 has over 35%aturated end groups and only for the less congestednd pentamethylzirconocenes (C5H5)(C5H5−nMen)ZrCl2n = 4, 5) chain transfer is less important (below 100%, respectively). This is attributed to steric cauhich make it more difficult for the�-hydrogen to geriented towards the metal center, so that chain traan become a major competing termination path81,82,102].

In a mechanistic study Siedle et al. uC5H5)2ZrMe2/MAO and 1-hexene oligomerizationoluene to find�-H elimination as the predominant charansfer which was different than for ethene oligomerizaith the same system (see above)[80].Hungenberg et al. investigated the oligomerization o

exene with (C5H5)2ZrCl2/MAO in bulk to establish a rataw which sufficiently allowed to describe time conversurves and the resulting oligomer distribution. Howehere was no dependence of the degree of polymerizr the oligomer distribution on conversion, i.e. on monooncentration[7].

The hydrodimerization of (E)- or (Z)-1-deuterio-1-hexeny achiral zirconocene dichloride/MAO [Eq. (16)] preseny Krauledat and Brintzinger[150] as well as the hydroyclization of trans,trans-1,6-d2-1,5-hexadiene by a scaocene hydride, Me2Si(C5Me4)2Sc(PMe3)H [Eq. (17)] car-ied out by Piers and Bercaw[151] gave results, whicere judged to provide experimental evidence for an�-gostic transition state (modified Green-Rooney mechan

n metal catalyzed olefin polymerizations. Theerythro:threo


or cis:trans ratio differing from unity in Eqs. (16) or(17), respectively, is seen to stem from an H/D-isotopeeffect, whose basis is the expected preference of Hover D in an �-agostic interaction for the transitionstates.

co-o tiona used( nd7

3

1 1-h utf(

3

iumc(({ ha tiono iums Eq.( eb tivei geds

{SiMe3}2)2Nd(�-Cl)2Li(THF)2 [111].

(18)

Ciardelli et al. carried out ethene/1-octene co-oligomerizations with (C5H5)2ZrMe2/MAO to investigatethe comonomer action as chain transfer agent (temperatures

ec-e

-

-],yl-q.ies

Ciardelli et al. carried out ethene/1-hexeneligomerizations to investigate the comonomer acs chain transfer agent. The co-oligomerizationsC5H5)2ZrMe2/MAO at two different temperatures (20 a0◦C), Al:Zr ratio = 1000[56].

.7. 1-Heptene oligomerization

Christoffers and Bergman activated (C5H5)2ZrCl2 by onlyequiv. of MAO to selectively and catalytically dimerizeeptene (and other�-olefins) to 2-pentyl-1-nonene witho

ormation of any higher oligomeric species [cf. Eq.(12) or14)] [51].

.8. 1-Octene oligomerization

Carpentier et al. showed that the trivalent neodymomplexes rac-Me2Si(C5H2-2,4-{SiMe3}2)2Nd(�-Cl)2Li-THF)2, rac-Me2Si(C5H2-3-SiMe3-4-tBu)2Nd(�-Cl)2Li-THF)2 and Me2Si(C5H2-2,4-{SiMe3}2)(C5H2-3,4-SiMe3}2)Nd(�-Cl)2Li(THF)2 when combined in situ witdialkylmagnesium cocatalyst, initiated the oligomerizaf 1-octene (and ethene) to yield di(oligoalkyl)magnespecies which were finally hydrolyzed to oligomers [18)] (Mn = 400–1300 g mol−1, Mw/Mn = 1.11–1.65). Thesulky bridged complexes were found to be more ac

n 1-octene oligomerization than the related nonbridystems (C5Me5)2Nd(�-Cl)2Li(OEt2)2 and (C5H3-1,3-

20 and 70◦C, Al:Zr ratio = 1000)[56].

3.9. Branched α-olefin oligomerization

Christoffers and Bergman reported that (C5H5)2ZrCl2when activated by only 1 equivalent of MAO seltively and catalytically dimerizes the�-olefins 1-buten(→ 2-ethyl-1-hexene) [Eq.(8)], 1-pentene (→ 2-propyl-1-heptene) [Eq.(12)], 1-hexene (→ 2-butyl-1-octene = 5methyleneundecane) [Eq.(14)], 1-heptene (→ 2-pentyl-1-nonene),3-methyl-1-butene (→ 5-methyl-2-(methylethyl)1-hexene = 2,6-dimethyl-3-methyleneheptane) [Eq. (19)4-methyl-1-pentene ([apparently wrongly named as 3-meth1-pentene]→ 6-methyl-2-(2-methylpropyl)-1-heptene) [E(20)] without formation of any higher oligomeric spec[51].


Longo et al. managed the selective dimerization of the�-branched�-olefin vinylcyclohexene in the presence of(C5H5)2MCl2 (M = Ti, Zr, Hf) or Me2Si(C5H4)2ZrCl2/MAO[Eq. (21)], with the latter one being by far the most ac-tive (8.9 g (mol l)Zr−1·(mol l)monomer−1 h−1). The latter cat-alyst was also used for the selective dimerization of the�-branched�-olefins vinylcyclopentene [Eq. (22)], 3-ethyl-1-pentene [Eq. (23)] and3,7-dimethyl-1-octene [Eq. (24)].Activities were even lower, however, with 3.5 g (mol l)Zr−1(mol l)monomer−1 h−1 for vinylcyclopentene being still thehighest. With3-methyl-1-pentene the product distributionwas 19% dimer, 50% trimer, 24% tetramer, 4% pentamer and3% of hexamer, with3-methyl-1-butene it was 11% dimer,36% trimer, 35% tetramer, 14% pentamer and 4% hexamer[152].

M o-a 1-h 7%od thet dc ,( q.(d -e e co-di

1o[

1 ec

(wi at− r inc

w ne,iat ain[

b ngt n-z dMt6

z ith(

Kretschmer et al. applied the dimeric complex{(2,4,7-e3-Ind′′′)2Y(�-H)}2 (19) as a catalyst (without any cctivator) in the regioselective homo-dimerization ofexene (→ 5-methyleneundecane, 80% yield plus 1f the trimer) [cf. Eq.(15)], 3-methyl-1-butene (→ 2,6-imethyl-3-methyleneheptane, 90% yield plus 9% of

rimer) [Eq. (25)], trimethylvinylsilane (→ head-to-heaoupling products (E)-1,4-bis(trimethylsilyl)but-1-ene, 48%Z)-∼ 24%, (E)-1,4-bis(trimethylsilyl)but-2-ene, 26%) [E26)], styrene (→ head-to-head coupling products (E)-1,4-iphenylbut-1-ene, 87%, (Z)-∼ 5%, (E)-1,3-diphenylbut-1ne, 6%) [Eq. (27)]. The same catalyst was also used in thimerization ofstyrene with �-olefins H2C CHR (R =nBu,

Pr, tBu, SiMe3, CH2Ph, CH2StBu and others) to givetrans--phenyl-4-alkyl-buten-1-enes, Ph-CHCH-CH2CH2-R inver 88% yield in benzene at 80–100◦C [Eq. (28)]149].

Ciardelli et al. co-oligomerizedethene with 4-methyl--pentene with (C5H5)2ZrMe2/MAO to investigate thomonomer action as chain transfer agent[56].

Shaffer and Ashbaugh oligomerizedisobuteneisobutylene) to a low Mn material (6600 g mol−1)ith Me2Si(Ind′H4)2ZrMe2/[Ph3C][B(C6F5)4] (1:1 ratio)

n the presence of∼2 equiv. of H2O in chlorobenzene20◦C. With considerably less amounts of residual wate

hlorobenzene highMn poly(isobutylene) is obtained[72].Hajela and Bercaw reacted Me2Si(C5Me4)2ScH(PMe3)

ith isobutene to observe the formation of 2-methylpentasobutane, 2-methyl-1-pentene, propane andn-pentane from

series of olefin insertion,�-Me and faster�-H elimina-ion which proceed until only the 2-methyl-1-alkenes rem153].

Kim et al. investigated the homopolymerization ofallyl-enzene with various metallocene/MAO catalysts. Amo

hem, (C5H5)2ZrCl2/MAO gave amorphous polyallylbeene with low molecular weight (Mw = 3400–1500 anw/Mn = 1.4–1.2 for a reaction temperature from−10

o 80◦C, respectively; highest activity at 60◦C with.8× 103 g molZr−1 h−1 [154].

Christoffers and Bergman dimerizedallylben-ene to 5-phenyl-2-(phenylmethyl)-1-pentene wC5H5)2ZrCl2/MAO (molar ratio 1:1) [Eq.(29)] [51].

(29)


Okamato et al. derived at optically active oligomers fromthe bulky allyltriphenylsilane in the reactions using opti-cally active C2H4(Ind′H4)2M(2,2′-biphenolate) or (binaph-tholate)/MAO catalysts (M = Zr, Hf)[155].

3.10. Cyclic olefin oligomerization

3.10.1. CyclopenteneCollins et al. demonstrated that the polymerization

of cyclopentene withrac-C2H4(Ind′)2ZrCl2/MAO leadsto the formation of poly(cis-1,3-cyclopentane) [Eq. (30)]through the independent synthesis of the single, stereoiso-meric tetramer under hydrooligomerization conditions[156].In contrast, hydrooligomerization of cyclopentene withrac-C2H4(Ind′H4)2ZrCl2/MAO led to the production ofoligomers in which cyclopentene is incorporated in acis- andtrans-1,3-manner [Eq. (31)] as was also confirmed throughindependent organic synthesis of some of these saturatedoligomers[92]. Steric reasons are seen responsible for hav-ing after the initial 1,2-insertion an isomerization (epimer-ization) through�-H elimination, followed by re-insertion ofthe olefin into the ZrH bond to give the 1,3-insertion prod-uct [73].

3e,

t ofHo fp ar-a nept2bb ctswn xtit

Xiwa

Tritto et al. studied the norbornene oligomerizationby (C5H5)2Ti(13Me)Cl/MAO in order to test the valid-ity of carbene mechanisms in�-olefin polymerizations. At0◦C reaction temperature, the addition product 2-(13C en-riched)methylnorbornane has been identified. At higher tem-peratures, the identification of a13C enriched methylidene-norbornane dimer revealed the possibility of norbornene ad-dition to Ti-carbenes through the formation of titanacyclobu-tane without the opening of the norbornene ring[159].

3.11. α,�-Diene oligomerization

Thiele and Erker used optimized reaction conditionsemploying low substrate concentrations (1.0–1.3 mol l−1,

imes-therith

(32)

.10.2. NorborneneFink et al. used the hydrooligomerization of norbornen

hat is the norbornene polymerization in the presence2, with rac-Me2C(Ind′)2ZrCl2/MAO to deliberately obtainligomers in order to gain insight into the microstructure oolynorbornene. The oligomers were separated by preptive HPLC and their structures, e.g. that of a norborneentamer (20), determined by single-crystal X-ray diffrac-

ion. These studies show that in-between the regularcis-,3-exo vinyl/addition insertions a metallocene-catalyzed�-ond metathesis can take place. Thesyn-hydrogen on C7 (theridgehead) of the previous to last inserted monomer interaith the Zr atom which in the�-bond metathesis becomesow bound to C7. Thus, the chain continues with the ne

nsertion on the C7-bridgehead carbon atom of the previouso last monomer[157].

Arndt and Gosmann had earlier described the pentamer-ray structure with the same bridgehead (C7) substitution

n a central trisubstituted norbornene unit (20). The pentameras isolated from a hydrooligomerization of norbornene cat-lyzed byrac-C2H4(Ind’H4)2ZrCl2/MAO [158].

hexadiene:Zr = 1900–2400) and rather long reaction t(5 h) to cyclodimerize1,5-hexadiene to give 1-methylene3-(cyclopentylmethyl)cyclopentane (36–58%) togewith trimers (30–36%) and tetramers (12–29%) w(C5H4R)2ZrCl2/MAO (R = H, Me, iPr, tBu; Al:Zr = 360;temperature 50–70◦C) [Eq.(32)] [54].

Christoffers and Bergman accomplished the oligomeriza-tion of 1,7-octadiene [Eq. (33)] with (C5H5)2ZrCl2/MAOto yield either lower oligomeric (1 d, RT,Mw = 1600,Mw/Mn = 2.0) or higher oligomeric materials (3 d, RT,Mw = 5400,Mw/Mn = 2.6)[160].


(33)

3.12. Alkyne oligomerization

The metallocene pre-catalyst (C5Me5)2AnMe2 (An = Th,U) promote the simultaneous production of a large numberof differently sized oligomers in the presence of terminalalkynes, HC CR (R = alkyl, aryl, SiMe3). The regioselectiv-ity and the extent of oligomerization depend strongly on thealkyne substituent R.

According to Eisen et al. reaction with HCCtBu yieldsregioselectively theE-2,4-bis(trimethylsilyl)-1-buten-3-ynehead-to-tail dimer [Eq. (34)] whereas HCCSiMe3 is regios-electively trimerized to the head-to-tail-to-head trimerE,E-1,4,6-tris(trimethylsilyl)-1,3-hexadiene-5-yne [Eq. (35)].Oligomerization with less bulky alkyl and aryl substitutedalkynes produces a mixture of higher oligomers with no re-gioselectivity[161,162].

darys horto inala

olledc ofd inala withtk thepa tion[

up 3c -l etala nt ofo = Y,a ead-t reso ynes

and 1,4-disubstituted 1-buten-3-ynes are found for HCCPhand HC CSiMe3 [Eq. (37)]. The reactions with Ln = La andCe produce not only dimers but also various higher oligomers(trimers, tetramers)[165].

Brubaker et al. attached titanocene dichloride to a poly-mer support via a covalent linkage to the cyclopentadienyl lig-and to form polymer-(C5H4)(C5H5)TiCl2 which on reductionwith excessnBuLi was found to oligomerize ethynyl propi-onate (=ethyl propiolate, HCC OC( O)Et) to a mixture ofopen and closed trimers. The closed trimers were identifiedas 1,2,4- and 1,3,5-tricarbethoxybenzene[166].

3h

un-u oly-m mer-i[

de-h -d -u es dro-c -s ane,P

SiHtH

A fb ea aseso lu-bM

Wang and Eisen described the addition of the seconilane Et2SiH2 to ensure the selective synthesis of sligomers, that is the selective dimerization of termlkynes promoted by (C5Me5)2UMe2 [163].

Eisen et al. also developed a strategy for the contratalytic synthesis (“tailoring”) of dimers or a mixtureimers and trimers through the oligomerization of termlkynes by using nonbulky or bulky amines, respectively,

he actinioid catalysts (C5Me5)2AnMe2 (An = Th, U). Theinetics in the catalytic oligomerization of 1-hexyne, inresence ofiBuNH2, by (C5Me5)2ThMe2 are first order inlkyne and Th and inverse first order in amine concentra

161,164].Heeres and Teuben investigated lanthanoid and gro

arbyls (C5Me5)2LnCH(SiMe3)2 (Ln = Y, La, Ce) as cataysts for the oligomerization of terminal alkynes. The mlso influenced strongly the regioselectivity and the exteligomerization besides the alkyne substitutent. For Lnlkyl-substituted alkynes are dimerized selectively to h

o-tail 2,4-disubstituted 1-buten-3-ynes [Eq. (36)]. Mixtuf two enyne isomers, e.g. 2,4-disubstituted 1-buten-3-

.13. Dehydrocoupling/dehydrooligomerization ofydrosilanes

Polysilane polymers are of interest because of theirsual electronic, optical and chemical properties. One perization method for these materials is the dehydropoly

zation of silanes with transition-metal catalysts [Eq.(38)]167].

(38)

Hilty et al. suggested the step-growth nature of theydrocoupling based on the (C5H5)2ZrMe2-catalyzed dehyropolymerization ofnBuSiH3, which produced low moleclar weight oligosilanes[168]. Tilley et al. illustrated thtep-growth character of the reaction by the slow dehyoupling of PhSiH3 by (C5Me5)2HfH2, which allows obervation of the early intermediates di-, tri- and tetrasilhH2Si-(–SiHPh–)n–SiH2Ph (n = 0, 1, 2)[167].Shaltout and Corey described the conversion of Ph3

o oligosilanes by rac- and meso-Me2C(C5H3SiMe3)2fCl2/nBuLi at room temperature[169].Woo et al. used (C5H5)2MCl2/(nBuLi or reduced

l) (M = Ti, Hf) for the dehydropolymerization ois(silyl)alkylbenzenes such as bis(1-sila-sec-butyl)benzennd 2-phenyl-1,3-disilapropane, producing two phf polymers. The molecular weights of the sole polymers ranged fromMn = 500–900 and fromw = 700–5000 g mol−1 (versus polystyrene)[170].


Dunogùes et al. synthesized poly(2,4-disilapentane)oligomers via dehydropolymerization of the monomerMeH2Si-CH2-SiMeH2 [Eq.(39)] using (C5H5)2ZrCl2/nBuLior (C5H5)2TiCl2/MeLi [171].

(39)

Tilley et al. proposed a mechanism of the dehydropoly-merization of hydrosilanes to polysilanes, as catalyzedby early-transition-metal metallocene derivatives. As partof their mechanistic studies they showed that the ther-molytic decomposition of (C5H5)(C5Me5)Hf(SiH2Ph)Cl to(C5H5)(C5Me5)Hf(H)Cl resulted in Si Si bond formationwith the production of polysilane oligomers[172].

Hengge and Weinberger obtained a mixture of theoligosilanes Me3Si(Me2Si)nSiMe2H from a dehydropoly-merization of pentamethyldisilane, Me3Si-SiHMe2 by meansof (C5H5)2MMe2 (M = Ti, Zr) (no co-activator neces-

Table 1Sandwich complexes with cyclic�-perimeters other than cyclopentadienyl for olefin oligomerization

Sandwich complex Monomer Remarks Ref.

Propene MAO cocatalyst; see also text in Section3.3 [132]

Propene, 1-hexene MAO cocatalyst; see also text in Sections3.3and 3.6

[81,82,102]

b
oratabenzene, C5H5BR− is a six �-electron aromaticanion, isoelectronic with benzene and resembling cy-clopentadienide
Ethene

1-octene

Ethene

1-decene

Propene

MAO cocatalyst;at 25◦C, 1 atm of C2H4; mixture of 1-alkenes,2-alkyl-1-alkenes and 2-alkenes; activities∼6.8–9.2× 105 g molzr−1 h−1.R = Ph, dimerization to 2-hexyl-1-decene

[178–181]

MAO cocatalyst;mixture of 1-alkenes, 2-alkyl-1-alkenes and2-alkenes; R = Et: 99%�-olefins; activities∼1.3–9.5× 105 g molzr−1 × h−1.R = Ph: decene dimerization to 2-octyl-1-dodecene

[179–183]

Complex synthesis from (C5Me5)MMe3(M = Zr, Hf) with neutral carborane C2B9H13

[184]


Table 2Half-sandwich complexes for olefin oligomerization

Half-sandwich complex Monomer Remarks Ref.

Propene, 1-butene, 1-pentene max.xi ≈ 70 for oligo-1-butene; chain transfer by�-H elimination

[186,187]

Ethene/styrene,ethene/p- or m-methylstyrene,ethene/p-tBu-styrene,ethene/p-chloro-styrene,1-hexene

Co-oligomerization mainly to 6-phenyl-1-hexene;subsequent co-polymerization of this oligomer withethene

[63,64]

[188]

Ethene

Ethene/1-hexene

MAO cocatalyst;mostly trimerization to 1-hexene as main product;study of various substituent effects on trimerizationproduct-% and productivity.co-trimerization to 5-methyl-non-1-ene

[189]

Pendant, hemi-labile arene group coordination to ac-tive center suggested:

(C5H5)TiCl3 immobilized on a cy-clopentadienyl surface of silica withnBuLi

Ethene MAO cocatalyst [190]

Ethene,propene,ethene/propene

AlEt2(OEt) cocatalyst; propene conversion tobranched dimers, mostly 2,3-dimethylbutenes

[191]

Ethene C4–C20 olefins;turnover > 105 mol(C2H4)·molcomplex−1 h−1

[192]


Table 2 (Continued )

Half-sandwich complex Monomer Remarks Ref.

1-HexeneMAO cocatalyst;activities between 2.6 and 5.6× 103 g molM−1 h−1

[193]

Ethene Activation with AliBu3/[Ph3C]- or [PhNHMe2]-[B(C6F5)4]

[113,114]

Ethene, propene1-pentene, 1-octene,4-methyl-1-pentene,4,4-dimethyl-1-pentene

Intermediate:

R = H, Me, nPr, nHex, CH2iPr, CH2tBu; selectivedimerization, mostly tail-to-tail dimers

[194]

Propene Mn = 2700 g mol−1, activity 1.4× 103 g gHf−1 h−1 atambient temperature and 5 bar

[195]

sary). Analogously 1,1,2,2-tetramethyldisilane, Me2HSi–SiHMe2 reacts to a mixture of the oligosilanes Me2HSi-(Me2Si)nSiHMe2 [173].

Kesti and Waymouth noted that when group-IV metal-locene catalysts, such as (C5H5)2ZrCl2/nBuLi, are used forthe hydrosilylation of olefins, silane coupling reactions com-pete with hydrosilylation in the case of less sterically hinderedsecondary or primary silanes and they produce oligomericproducts from dehydropolymerization[174].

4. Complexes related to metallocenes for olefinoligomerization

4.1. Sandwich complexes with cyclic π-perimeters otherthan cyclopentadienyl

The term “metallocene” refers to compounds which havea metal atom bound to two cyclopentadienide anions in a�-

or pentahapto/�5-mode. The cyclopentadienyl rings can co-ordinate in a co-planar (parallel) or bent fashion to the metalcenter. Besides the two cyclopentadienyl ring

Documents

Review Metallocene and related catalysts for oleﬁn, alkyne …...Review Metallocene and related catalysts for oleﬁn, alkyne and silane dimerization and oligomerization Christoph